- Email: [email protected]
The effects of gas-fluid-rock interactions on CO2 injection and storage: Insights from reactive transport modeling
Available online at www.sciencedirect.com
Energy Procedia
Energy (2009) EnergyProcedia Procedia1 00 (2008)1783–1790 000–000 www.elsevier.com/locate/procedia www.elsevier.com/locate/XXX
GHGT-9
The effects of gas-fluid-rock interactions on CO2 injection and storage: insights from reactive transport modeling Yitian Xiaoa*, Tianfu Xub, and Karsten Pruessb a
ExxonMobil Upstream Research Company, Houston TX 77027, USA Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
b
Elsevier use only: Received date here; revised date here; accepted date here
Abstract Possible means of reducing atmospheric CO2 emissions include injecting CO2 in petroleum reservoirs for Enhanced Oil Recovery or storing CO2 in deep saline aquifers. Large-scale injection of CO2 into subsurface reservoirs would induce a complex interplay of multiphase flow, capillary trapping, dissolution, diffusion, convection, and chemical reactions that may have significant impacts on both short-term injection performance and long-term fate of CO2 storage. Reactive Transport Modeling is a promising approach that can be used to predict the spatial and temporal evolution of injected CO2 and associated gas-fluid-rock interactions. This presentation will summarize recent advances in reactive transport modeling of CO2 storage and review key technical issues on (1) the short- and long-term behavior of injected CO2 in geological formations; (2) the role of reservoir mineral heterogeneity on injection performance and storage security; (3) the effect of gas mixtures (e.g., H2S and SO2) on CO2 storage; and (4) the physical and chemical processes during potential leakage of CO2 from the primary storage reservoir. Simulation results suggest that CO2 trapping capacity, rate, and impact on reservoir rocks depend on primary mineral composition and injecting gas mixtures. For example, models predict that the injection of CO2 alone or co-injection with H2S in both sandstone and carbonate reservoirs lead to acidified zones and mineral dissolution adjacent to the injection well, and carbonate precipitation and mineral trapping away from the well. Co-injection of CO2 with H2S and in particular with SO2 causes greater formation alteration and complex sulfur mineral (alunite, anhydrite, and pyrite) trapping, sometimes at a much faster rate than previously thought. The results from Reactive Transport Modeling provide valuable insights for analyzing and assessing the dynamic behaviors of injected CO2, identifying and characterizing potential storage sites, and managing injection performance and reducing costs. c© 2009
2008 Elsevier Elsevier Ltd. Ltd. All All rights rights reserved. reserved Keywords: CO2 injection and storage, gas-fluid-rock interactions, reactive transport modeling, ccs.
1. Introduction Possible means of reducing atmospheric CO2 emissions include injecting CO2 in petroleum reservoirs for EOR or storing CO2 in deep saline aquifers. Sequestering raw CO2 containing H2S and/or SO2 requires less energy to separate from flue gas or a coal gasification process and therefore might be the preferred disposal option [1, 2]. Large-scale injection of CO2 and other gases into subsurface reservoirs may induce a complex interplay of multiphase flow, dissolution, precipitation, diffusion, convection, and other chemical reactions [3, 4]. Compared to CO2 residual and solubility trapping, mineral trapping of CO2 is potentially attractive because it could immobilize
doi:10.1016/j.egypro.2009.01.233
1784
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
2
Xiao et al / Energy Procedia 00 (2008) 000–000
CO2 for long time scales and increase storage security. Depending on the spatial distribution and temporal evolution of the CO2 injection and associated mineral dissolution and precipitation (artificial diagenesis), the gas-fluid-rock interactions could have a significant impact on injection performance and storage capacity and security. Alteration of the predominant host rock minerals is usually very slow and therefore is not experimentally accessible under deep reservoir or aquifer conditions. Reactive Transport Modeling (RTM) is a promising approach that can be used to predict the spatial and temporal evolution of injected CO2/H2S/SO2 and associated formation alteration [1, 2]. The objective of this study was to conduct reactive transport simulations to investigate mineral alteration and sequestration of mixed CO2, H2S and SO2 in a siliciclastic and a carbonate reservoir. 2. Reactive Transport Modeling Approach A reactive transport model relies on a mathematical formulation to describe geochemical processes involving fluid-rock interactions. The general governing equation can be written as: w (ICi ) wt
wC w § wCi · § wC · ¸ Iv i I ¦ ¨ i ¸ ¨ ID wx wx © wx ¹ k © wt ¹ k
(1)
Where Ci is the concentration of a specific species in the pore fluid, D is the combined diffusion and dispersion coefficient term, v is the linear fluid flow rate, and ɮ is the porosity. The first two terms on the right hand side describe the transport process (diffusion, dispersion, and advection) while the last term describes the effect of geochemical reactions (Figure 1).
Simulation tracks fluid flow and chemical reactions in porous media
Inflow
dissolution/precipitation with changes in porosity/permeability
Mg2+
Outflow
Ca2+
Initial rock property
Dolomitization: 2CaCO 3 + Mg 2+ = CaMg(CO 3)2 + Ca 2+
Figure 1. Schematic diagram showing the reactive transport process involving dolomitization.
For example, the reaction of dolomitization can be expressed as:
wX dolo wt
S Ae
Ea RT
(
Q 1) 2.26 K eq
(2)
Where Xdolo is the dolomite fraction, S represents the reactive surface area, A is the rate constant, Ea is the activation energy, Q/Keq is the saturation index, and 2.26 is the reaction order [5]. Due to complex boundary conditions and complicated coupling between the transport and reaction terms, it is impossible to provide analytical solutions to equation (1) for even the simplest geochemical system. Therefore numerical solutions have to be used. Fortunately, due to the exponential increase in computational power, realistic reactive transport models are beginning to provide new insights to CO2 injection and storage at both injection and geological time scales. The simulations in this study were carried out using the non-isothermal reactive geochemical transport code TOUGHREACT [6]. The program treats multi-phase fluid and heat flow, advection and diffusion. It models geochemical reactions including aqueous complexation, mineral dissolution and precipitation, gas
1785
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
3
dissolution and exsolution, and ion exchange. Special modeling considerations include CO2 solubility dependence on P, T, and salinity, changes in porosity and permeability due to mineral dissolution and precipitation, gas phase and gaseous species are active in flow, transport, and other reactions. Two previous studies [1, 2] applied reactive transport modeling to investigate mixed CO2/H2S/SO2 injections in sandstone reservoirs, this study covered gas injection and storage in both siliciclastic and carbonate reservoirs. 3. Reactive Transport Modeling Design The 1D radial reactive transport models represent CO2 injection in a siliciclastic and carbonate reservoir at 2 km depth and 70 oC. CO2 and other gases were injected in the reservoir at a rate of 1 million ton per year over a period of 100 years. The reactive transport models simulate the system from 0 to 10,000 years. There are three scenarios of mixed gas injected: CO2 only, CO2 + H2S, and CO2 + SO2 in which CO2 is injected as gas phase while both H2S and SO2 (~5% each) are injected as aqueous solutes. The reservoirs are specified to have an initial porosity of 0.30 and initial permeability of 100 mD. The siliciclastic and carbonate reservoirs were defined by hypothetical mineral assemblages, representing an oligoclase/feldspar-rich sandstone reservoir and a limestone-rich reservoir, respectively (Table 1). Other primary and secondary minerals are listed in Table 1 as well. Table 1. Initial mineral compositions of the siliciclastic and carbonate reservoirs and secondary minerals considered in the Simulation (after Xu et al [1]). Mineral
Primary: Quartz Kaolinite Calcite Illite Oligoclase K-feldspar Na-smectite Chlorite Hematite Porosity Secondary: Anhydrite Magnesite Dolomite Low-albite Siderite Ankerite Dawsonite Ca-smectite Alunite Pyrite Opal-A
Chemical formula
Volume percent Siliciclastic Carbonate Reservoir Reservoir
SiO2 Al2Si2O5(OH)4 CaCO3 K0.6Mg0.25Al1.8(Al0.5Si3.5O10)(OH)2 Ca0.2Na0.8Al1.2Si2.8O8 KAlSi3O8 Na0.290Mg0.26Al1.77Si3.97O10(OH)2 Mg2.5 Fe2.5Al2Si3O10(OH)8 Fe2O3
40.6 1.41 1.35 0.7 13.86 5.74 4.8 1.19 0.35 30
CaSO4 MgCO3 CaMg(CO3)2 NaAlSi3O8 FeCO3 CaMg0.3Fe0.7(CO3)2 NaAlCO3 (OH) 2 Na0.145Mg0.26Al1.77Si3.97O10(OH)2 KAl3 (OH) 6(SO4) 2 FeS2 SiO2
1.0 1.5 63.0 0.6 0.5 1.2 0.6 1.6 0.00 30
1786
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
4
Xiao et al / Energy Procedia 00 (2008) 000–000
4. Simulation Results: Sandstone Reservoir Injection 4.1 pH Evolution
9.0
9.0
8.0
8.0
8.0
7.0
7.0
7.0
6.0
6.0
6.0
5.0
5.0
5.0
4.0
10 yr
3.0
100 yr
2.0
1,000 yr
1.0
10,000 yr
0.0 0
1
10
100
1000
10000
pH
9.0
pH
pH
The simulation results from the CO2 only case are similar to those from CO2 + H2S. The results from CO2 + SO2 case are however, significantly different (Figure 2). CO2 and CO2 + H2S injection leads to lower pH (~4) near the well bore, resulting from the dissolution of CO2 and H2S in the formation water. Once the injection stops, the pH is quickly buffered by the reservoir minerals and goes back to a more neutral value (~6). Co-injection of SO2 with CO2 leads to very low pH (~0), presumably due to the dissolution of SO2 and the formation of sulphuric acid zone close to the injection well. Corrosion and well abandonment are potential issues. After injection, the pH is buffered but still remains considerably low near the well bore.
4.0
4.0
3.0
3.0
2.0
2.0
1.0
1.0
0.0
0.0
0
1
10
100
Radial Distance (m)
Radial Distance (m)
CO2 Only
CO2 + H2S
1000
pH ~ 0! 0
10000
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 2. Simulated pH distribution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
4.2 Mineral Dissolution/Precipitation Simulation results reveal that both CO2 and CO2 + H2S injections would dissolve oligoclase and precipitate smectite, although at a relatively slow rate (~0.1%/yr) with a gradual reaction front extending 100 meters. The CO2 + SO2 injection however, dissolves both minerals at much faster rates (~1%/yr) with a sharp reaction front (Figure 3).
1000 yr
10 yr
0.12
- Oligoclase
0.10 0.08
100 yr
- Smectite 100 yr
0.06 0.04
10 yr
0.02
1000 yr
0.00
0.16
0.16
0.14
0.14
Oligoclase-Smectite (vol %)
0.14
Oligoclase-Smectite (vol %)
Oligoclase-Smectite (vol %)
0.16
0.12 0.10 0.08
- Oligoclase dissolution - Smectite precipitation
0.06 0.04 0.02 0.00
0
1
10
100
1000
Radial Distance (m)
CO2 Only
10000
Oligoclase dissolution
0.12
Smectite precipitation
0.10 0.08 0.06 0.04 0.02 0.00
0
1
10
100
1000
Radial Distance (m)
CO2 + H2S
10000
0
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 3. Simulated mineral (oligoclase and smectite) evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir.
1787
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
5
4.3 CO2 Sequestration Capacity Simulation results indicate that most sequestered CO2 is in the form of ankerite and dawsonite, and can reach 80 kg/m3 or ~5% of bulk mineral weight over 10,000 years (Figure 4). The mineral trapping zones associated with CO2 and CO2 + H2S injection are broader while CO2 + SO2 injection leads to a narrower trapping zone farther away from the injection well. The RTM results are consistent with field observations [7, 8] and natural analogue studies [9].
100 yr
60
1,000 yr
50
10,000 yr
40 30 20 10
80
80
70
70
Dawsonite NaAlCO3 (OH) 2 60 Ankerite CaMg0.3Fe0.7(CO3)2
CO2 Sequestered (Kg/m**3)
10 yr
70
CO2 Sequestered (Kg/m**3)
CO2 Sequestered (Kg/m**3)
80
50 40 30 20 10
60 50 40 30 20 10 0
0
0 0
1
10
100
1000
0
10000
Dawsonite NaAlCO3 (OH) 2 Ankerite CaMg0.3Fe0.7(CO3)2
1
10
100
1000
0
10000
10
CO2 Only
1000
Radial Distance (m)
Radial Distance (m)
Radial Distance (m)
CO2 + H2S
CO2 + SO2
Figure 4. Simulated mineral trapping capacity (via dawsonite and ankerite) as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
5. Simulation Results: Carbonate Reservoir Injection 5.1 pH Evolution The simulation results from the CO2 only case are similar to those from CO2 + H2S. The results from CO2 + SO2 case are however, significantly different (Figure 5). CO2 and CO2 + H2S injection leads to lower pH (~4.5, less so comparing to siliciclastic reservoir) near the well bore, resulting from the dissolution of CO2 and H2S in the formation water and quicker buffering reactions. Once the injection stops, the pH increases slightly (~5) but does not rebound to more neutral pH due to limited buffering capability. Co-injection of SO2 with CO2 leads to very low pH (~0), presumably due to the dissolution of SO2 and the formation of a sulphuric acid zone close to the injection well. Corrosion and well abandonment can be potential issues. After the injection, the pH is buffered but still remains considerably low near the well bore. 8.0
8.0
7.5
7.0
1,000 yr
7.0
6.0
10,000 yr
6.5
5.0
100 yr 7.0
6.0
pH
pH
6.5
pH
8.0
10 yr
7.5
6.0
4.0
5.5
5.5
3.0
5.0
5.0
2.0
4.5
4.5
1.0
4.0
4.0
0
1
10
100
Radial Distance (m)
CO2 Only
1000
10000
pH ~ 0!
0.0 0
1
10
100
Radial Distance (m)
CO2 + H2S
1000
10000
0
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 5. Simulated pH distribution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
1788
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
6
Xiao et al / Energy Procedia 00 (2008) 000–000
5.2 Mineral Dissolution/Precipitation Simulation results reveal that calcite dissolution occurs in all cases at a much faster rate compared to the siliciclastic reservoir injection. Under mixed CO2 + SO2 injection, significant and rapid calcite dissolution and anhydrite precipitation occurs near the well bore with a sharp reaction front. The latter could lead to significant loss of injection performance due to porosity reduction. The simulation results are comparable with field observations of acid gas injections in carbonate reservoirs [10]. 0.8
0.8
0.7
0.7
0 yr
0.4
10 yr
0.3
100 yr
0.2
1,000 yr 10,000 yr
0.1 0.0
0.6
Calcite (vol %)
Calcite (vol %)
0.5
0.5 0.4
Calcite Dissolution
0.3
1
10
100
1000
0.4 0.3 0.2
0.1
0.1
10000
0
Radial Distance (m)
1
10
100
1000
10000
0.0 0
Radial Distance (m)
CO2 Only
0 yr
0.5
0.2
0.0 0
Calcite Dissolution Anhydrite Precipitation
0.7
0 yr
0.6
0.6
Calcite (vol %)
0.8
1
10
100
1000
10000
Radial Distance (m)
CO2 + H2S
CO2 + SO2
Figure 6. Simulated mineral (calcite) evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
6. Porosity Evolution Siliciclastic Reservoir: The gas injection leads to an increase in porosity close to the well due to net mineral dissolution, and a decrease at distances due to mineral trapping in all three cases (Figure 7). However, the porosity gains associated with CO2 and CO2 + H2S injection are much smaller (0.30 to 0.32) compared to the CO2 + SO2 injection (0.30 to 0.50) at 100 years. The same trend is observed for mineral trapping. 0.6
10 yr
0.6
0.6
0.5
0.5
Dissolution
1,000 yr 10,000 yr
0.4
Dissolution
0.3
Porosity
Porosity
0.5
0.4
Dissolution
Trapping 0.2
0
10 Radial Distance (m)
CO2 Only
1000
Trapping 0 yr
0 yr
Trapping 0.2
0.4
0.3
0.3
0 yr
Porosity
100 yr
0
10 Radial Distance (m)
CO2 + H2S
1000
0.2 0
10
1000
Radial Distance (m)
CO2 + SO2
Figure 7. Simulated porosity evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
Carbonate Reservoir: Simulation results indicate that there are significant porosity increases (0.30 to 0.40 at 100 years) associated with CO2 and CO2 + H2S injection due to calcite dissolution near the well bore (Figure 8).
1789
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
7
However, there is significant porosity decrease (0.30 to 0.20) associated with CO2 + SO2 injection due to anhydrite precipitation. There is little mineral trapping in all three mixed gas injection scenarios, suggesting limited CO2 sequestration capacity in a limestone dominated carbonate reservoir. 0.45 0.40
Calcite Dissolution
0.40
0.35
0.25
0 yr
100 yr 1,000 yr
0.20
0.30
0 yr
0
10 1000 Radial Distance (m)
CO2 Only
0.30
0 yr
0.25
0.25
0.20
0.20
0.15
0.15
10,000 yr 0.15
Porosity
10 yr
Porosity
0.30
Calcite Dissolution Anhydrite Precipitation
0.40
0.35
0.35 Porosity
0.45
0.45
Calcite Dissolution
0
10 1000 Radial Distance (m)
CO2 + H2S
Porosity Decrease 0
10 1000 Radial Distance (m)
CO2 + SO2
Figure 8. Simulated porosity evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
7. Summary and Conclusions x x x x
x
The behavior of CO2 and acid gas injection and storage is controlled by gas mixtures, reservoir mineralogy, timing, and injection design. Co-injection of H2S yields similar behavior compared to CO2 injection. Co-injection of SO2 results in the formation of sulfuric acid close to the well. Corrosion of pipes and well abandonment are potential issues. Significantly more CO2 is sequestered as ankerite and dawsonite in siliciclastic reservoirs than in carbonate reservoirs, with the mineral trapping capability reaching as much as 80 kg/m3. Most injection scenarios result in porosity increase close to the well and decrease at distances. However, co-injection of SO2 in a carbonate reservoir leads to significant anhydrite precipitation and porosity reduction in the near-well region. The results from Reactive Transport Modeling provide valuable insights for describing, analyzing, interpreting, and assessing the physical properties and dynamic behaviours of injected CO2 and facilitating the screening and evaluation of CO2 storage strategy.
8. Acknowledgments The paper benefited from fruitful discussions with ExxonMobil colleagues Will Maze, Gary Teletzke, John Wilkinson, Lee Esch, Bob Pottorf, and Mark Richardson. Careful reviews by Chris Reaves and Bill Clendenen improved the manuscript. Permission by ExxonMobil to release the paper is greatly appreciated. Drs. Tianfu Xu and Karsten Pruess were supported by the Zero Emission Research and Technology project (ZERT) of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 with Lawrence Berkeley National Laboratory.
1790
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
8
Xiao et al / Energy Procedia 00 (2008) 000–000
9. References 1.
Xu, T, J. A. Apps, K. Pruess, and H. Yamamoto, 2007. Numerical modeling of injection and mineral trapping of CO2 with H2S and SO2 in a sandstone formation. Chemical Geology, Volume 242, Issues 3-4, 15 August 2007, Pages 319-346
2.
Knauss, K.G., Johnson, J.W., Steefel, C.I., 2005. Evaluation of the impact of CO2, co-contaminant gas, aqueous fluid, and reservoir rock interactions on the geologic sequestration of CO2. Chem. Geol. 217, 339350.
3.
Pruess, K., Xu, T., Apps, J., García, J., 2003. Numerical modeling of aquifer disposal of CO2. SPE (Society of Petroleum Engineers) Journal, p. 49-59, March (2003).
4.
Johnson, J.W., Nitao, J.J., Steefel, C.I., Knaus, K.G., 2001. Reactive transport modeling of geologic CO2 sequestration in saline aquifers: The influence of intra-aquifer shales and the relative effectiveness of structural, solubility, and mineral trapping during prograde and retrograde sequestration, In proceedings: First National Conference on Carbon Sequestration. Washington, DC.
5.
Arvidson, R. S. and Mackenzie, F.T. 1999. The dolomite problem: control of precipitation kinetics by temperature and saturation state: American Journal of Science, v. 299, p. 257-288.
6.
Xu, T., E.L. Sonnenthal, N. Spycher, and K. Pruess, 2006. TOURGHREACT: A simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media. Computer & Geosciences 32, 145-165.
7.
Watson, M.N., Zwingmann, N., Lemon, N.M., 2004. The Ladbroke Grove–Katnook carbon dioxide natural laboratory: A recent CO2 accumulation in a lithic sandstone reservoir. Energy, 29(9-10), 1457-1466.
8.
Moore, J., Adams, M., Allis, R., Lutz, S., Rauzi, S., 2005. Mineralogical and geochemical consequences of the long-term presence of CO2 in natural reservoirs: An example from the Springerville–St. Johns Field, Arizona, and New Mexico, U.S.A. Chemical Geology, 217, 365-385.
9.
Worden R H (2006) Dawsonite cement in the Triassic Lam Formation, Shabwa Basin, Yemen: a natural analogue for a potential mineral product of subsurface CO2 storage for greenhouse gas reduction. Marine and Petroleum Geology, 23, pp.61-67.
10. Huang, N S, G.E. Aho, B.H. Baker and P.A. Foglesong, T.R. Matthews, R.J. Pottorf, 2007. Integrated Reservoir Modeling to Maximize the Value of a Large Sour-Gas Field with High Concentrations of Inerts. 2007 International Petroleum Technology Conference, Dubai, UAE.
Energy Procedia
Energy (2009) EnergyProcedia Procedia1 00 (2008)1783–1790 000–000 www.elsevier.com/locate/procedia www.elsevier.com/locate/XXX
GHGT-9
The effects of gas-fluid-rock interactions on CO2 injection and storage: insights from reactive transport modeling Yitian Xiaoa*, Tianfu Xub, and Karsten Pruessb a
ExxonMobil Upstream Research Company, Houston TX 77027, USA Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA
b
Elsevier use only: Received date here; revised date here; accepted date here
Abstract Possible means of reducing atmospheric CO2 emissions include injecting CO2 in petroleum reservoirs for Enhanced Oil Recovery or storing CO2 in deep saline aquifers. Large-scale injection of CO2 into subsurface reservoirs would induce a complex interplay of multiphase flow, capillary trapping, dissolution, diffusion, convection, and chemical reactions that may have significant impacts on both short-term injection performance and long-term fate of CO2 storage. Reactive Transport Modeling is a promising approach that can be used to predict the spatial and temporal evolution of injected CO2 and associated gas-fluid-rock interactions. This presentation will summarize recent advances in reactive transport modeling of CO2 storage and review key technical issues on (1) the short- and long-term behavior of injected CO2 in geological formations; (2) the role of reservoir mineral heterogeneity on injection performance and storage security; (3) the effect of gas mixtures (e.g., H2S and SO2) on CO2 storage; and (4) the physical and chemical processes during potential leakage of CO2 from the primary storage reservoir. Simulation results suggest that CO2 trapping capacity, rate, and impact on reservoir rocks depend on primary mineral composition and injecting gas mixtures. For example, models predict that the injection of CO2 alone or co-injection with H2S in both sandstone and carbonate reservoirs lead to acidified zones and mineral dissolution adjacent to the injection well, and carbonate precipitation and mineral trapping away from the well. Co-injection of CO2 with H2S and in particular with SO2 causes greater formation alteration and complex sulfur mineral (alunite, anhydrite, and pyrite) trapping, sometimes at a much faster rate than previously thought. The results from Reactive Transport Modeling provide valuable insights for analyzing and assessing the dynamic behaviors of injected CO2, identifying and characterizing potential storage sites, and managing injection performance and reducing costs. c© 2009
2008 Elsevier Elsevier Ltd. Ltd. All All rights rights reserved. reserved Keywords: CO2 injection and storage, gas-fluid-rock interactions, reactive transport modeling, ccs.
1. Introduction Possible means of reducing atmospheric CO2 emissions include injecting CO2 in petroleum reservoirs for EOR or storing CO2 in deep saline aquifers. Sequestering raw CO2 containing H2S and/or SO2 requires less energy to separate from flue gas or a coal gasification process and therefore might be the preferred disposal option [1, 2]. Large-scale injection of CO2 and other gases into subsurface reservoirs may induce a complex interplay of multiphase flow, dissolution, precipitation, diffusion, convection, and other chemical reactions [3, 4]. Compared to CO2 residual and solubility trapping, mineral trapping of CO2 is potentially attractive because it could immobilize
doi:10.1016/j.egypro.2009.01.233
1784
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
2
Xiao et al / Energy Procedia 00 (2008) 000–000
CO2 for long time scales and increase storage security. Depending on the spatial distribution and temporal evolution of the CO2 injection and associated mineral dissolution and precipitation (artificial diagenesis), the gas-fluid-rock interactions could have a significant impact on injection performance and storage capacity and security. Alteration of the predominant host rock minerals is usually very slow and therefore is not experimentally accessible under deep reservoir or aquifer conditions. Reactive Transport Modeling (RTM) is a promising approach that can be used to predict the spatial and temporal evolution of injected CO2/H2S/SO2 and associated formation alteration [1, 2]. The objective of this study was to conduct reactive transport simulations to investigate mineral alteration and sequestration of mixed CO2, H2S and SO2 in a siliciclastic and a carbonate reservoir. 2. Reactive Transport Modeling Approach A reactive transport model relies on a mathematical formulation to describe geochemical processes involving fluid-rock interactions. The general governing equation can be written as: w (ICi ) wt
wC w § wCi · § wC · ¸ Iv i I ¦ ¨ i ¸ ¨ ID wx wx © wx ¹ k © wt ¹ k
(1)
Where Ci is the concentration of a specific species in the pore fluid, D is the combined diffusion and dispersion coefficient term, v is the linear fluid flow rate, and ɮ is the porosity. The first two terms on the right hand side describe the transport process (diffusion, dispersion, and advection) while the last term describes the effect of geochemical reactions (Figure 1).
Simulation tracks fluid flow and chemical reactions in porous media
Inflow
dissolution/precipitation with changes in porosity/permeability
Mg2+
Outflow
Ca2+
Initial rock property
Dolomitization: 2CaCO 3 + Mg 2+ = CaMg(CO 3)2 + Ca 2+
Figure 1. Schematic diagram showing the reactive transport process involving dolomitization.
For example, the reaction of dolomitization can be expressed as:
wX dolo wt
S Ae
Ea RT
(
Q 1) 2.26 K eq
(2)
Where Xdolo is the dolomite fraction, S represents the reactive surface area, A is the rate constant, Ea is the activation energy, Q/Keq is the saturation index, and 2.26 is the reaction order [5]. Due to complex boundary conditions and complicated coupling between the transport and reaction terms, it is impossible to provide analytical solutions to equation (1) for even the simplest geochemical system. Therefore numerical solutions have to be used. Fortunately, due to the exponential increase in computational power, realistic reactive transport models are beginning to provide new insights to CO2 injection and storage at both injection and geological time scales. The simulations in this study were carried out using the non-isothermal reactive geochemical transport code TOUGHREACT [6]. The program treats multi-phase fluid and heat flow, advection and diffusion. It models geochemical reactions including aqueous complexation, mineral dissolution and precipitation, gas
1785
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
3
dissolution and exsolution, and ion exchange. Special modeling considerations include CO2 solubility dependence on P, T, and salinity, changes in porosity and permeability due to mineral dissolution and precipitation, gas phase and gaseous species are active in flow, transport, and other reactions. Two previous studies [1, 2] applied reactive transport modeling to investigate mixed CO2/H2S/SO2 injections in sandstone reservoirs, this study covered gas injection and storage in both siliciclastic and carbonate reservoirs. 3. Reactive Transport Modeling Design The 1D radial reactive transport models represent CO2 injection in a siliciclastic and carbonate reservoir at 2 km depth and 70 oC. CO2 and other gases were injected in the reservoir at a rate of 1 million ton per year over a period of 100 years. The reactive transport models simulate the system from 0 to 10,000 years. There are three scenarios of mixed gas injected: CO2 only, CO2 + H2S, and CO2 + SO2 in which CO2 is injected as gas phase while both H2S and SO2 (~5% each) are injected as aqueous solutes. The reservoirs are specified to have an initial porosity of 0.30 and initial permeability of 100 mD. The siliciclastic and carbonate reservoirs were defined by hypothetical mineral assemblages, representing an oligoclase/feldspar-rich sandstone reservoir and a limestone-rich reservoir, respectively (Table 1). Other primary and secondary minerals are listed in Table 1 as well. Table 1. Initial mineral compositions of the siliciclastic and carbonate reservoirs and secondary minerals considered in the Simulation (after Xu et al [1]). Mineral
Primary: Quartz Kaolinite Calcite Illite Oligoclase K-feldspar Na-smectite Chlorite Hematite Porosity Secondary: Anhydrite Magnesite Dolomite Low-albite Siderite Ankerite Dawsonite Ca-smectite Alunite Pyrite Opal-A
Chemical formula
Volume percent Siliciclastic Carbonate Reservoir Reservoir
SiO2 Al2Si2O5(OH)4 CaCO3 K0.6Mg0.25Al1.8(Al0.5Si3.5O10)(OH)2 Ca0.2Na0.8Al1.2Si2.8O8 KAlSi3O8 Na0.290Mg0.26Al1.77Si3.97O10(OH)2 Mg2.5 Fe2.5Al2Si3O10(OH)8 Fe2O3
40.6 1.41 1.35 0.7 13.86 5.74 4.8 1.19 0.35 30
CaSO4 MgCO3 CaMg(CO3)2 NaAlSi3O8 FeCO3 CaMg0.3Fe0.7(CO3)2 NaAlCO3 (OH) 2 Na0.145Mg0.26Al1.77Si3.97O10(OH)2 KAl3 (OH) 6(SO4) 2 FeS2 SiO2
1.0 1.5 63.0 0.6 0.5 1.2 0.6 1.6 0.00 30
1786
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
4
Xiao et al / Energy Procedia 00 (2008) 000–000
4. Simulation Results: Sandstone Reservoir Injection 4.1 pH Evolution
9.0
9.0
8.0
8.0
8.0
7.0
7.0
7.0
6.0
6.0
6.0
5.0
5.0
5.0
4.0
10 yr
3.0
100 yr
2.0
1,000 yr
1.0
10,000 yr
0.0 0
1
10
100
1000
10000
pH
9.0
pH
pH
The simulation results from the CO2 only case are similar to those from CO2 + H2S. The results from CO2 + SO2 case are however, significantly different (Figure 2). CO2 and CO2 + H2S injection leads to lower pH (~4) near the well bore, resulting from the dissolution of CO2 and H2S in the formation water. Once the injection stops, the pH is quickly buffered by the reservoir minerals and goes back to a more neutral value (~6). Co-injection of SO2 with CO2 leads to very low pH (~0), presumably due to the dissolution of SO2 and the formation of sulphuric acid zone close to the injection well. Corrosion and well abandonment are potential issues. After injection, the pH is buffered but still remains considerably low near the well bore.
4.0
4.0
3.0
3.0
2.0
2.0
1.0
1.0
0.0
0.0
0
1
10
100
Radial Distance (m)
Radial Distance (m)
CO2 Only
CO2 + H2S
1000
pH ~ 0! 0
10000
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 2. Simulated pH distribution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
4.2 Mineral Dissolution/Precipitation Simulation results reveal that both CO2 and CO2 + H2S injections would dissolve oligoclase and precipitate smectite, although at a relatively slow rate (~0.1%/yr) with a gradual reaction front extending 100 meters. The CO2 + SO2 injection however, dissolves both minerals at much faster rates (~1%/yr) with a sharp reaction front (Figure 3).
1000 yr
10 yr
0.12
- Oligoclase
0.10 0.08
100 yr
- Smectite 100 yr
0.06 0.04
10 yr
0.02
1000 yr
0.00
0.16
0.16
0.14
0.14
Oligoclase-Smectite (vol %)
0.14
Oligoclase-Smectite (vol %)
Oligoclase-Smectite (vol %)
0.16
0.12 0.10 0.08
- Oligoclase dissolution - Smectite precipitation
0.06 0.04 0.02 0.00
0
1
10
100
1000
Radial Distance (m)
CO2 Only
10000
Oligoclase dissolution
0.12
Smectite precipitation
0.10 0.08 0.06 0.04 0.02 0.00
0
1
10
100
1000
Radial Distance (m)
CO2 + H2S
10000
0
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 3. Simulated mineral (oligoclase and smectite) evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir.
1787
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
5
4.3 CO2 Sequestration Capacity Simulation results indicate that most sequestered CO2 is in the form of ankerite and dawsonite, and can reach 80 kg/m3 or ~5% of bulk mineral weight over 10,000 years (Figure 4). The mineral trapping zones associated with CO2 and CO2 + H2S injection are broader while CO2 + SO2 injection leads to a narrower trapping zone farther away from the injection well. The RTM results are consistent with field observations [7, 8] and natural analogue studies [9].
100 yr
60
1,000 yr
50
10,000 yr
40 30 20 10
80
80
70
70
Dawsonite NaAlCO3 (OH) 2 60 Ankerite CaMg0.3Fe0.7(CO3)2
CO2 Sequestered (Kg/m**3)
10 yr
70
CO2 Sequestered (Kg/m**3)
CO2 Sequestered (Kg/m**3)
80
50 40 30 20 10
60 50 40 30 20 10 0
0
0 0
1
10
100
1000
0
10000
Dawsonite NaAlCO3 (OH) 2 Ankerite CaMg0.3Fe0.7(CO3)2
1
10
100
1000
0
10000
10
CO2 Only
1000
Radial Distance (m)
Radial Distance (m)
Radial Distance (m)
CO2 + H2S
CO2 + SO2
Figure 4. Simulated mineral trapping capacity (via dawsonite and ankerite) as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
5. Simulation Results: Carbonate Reservoir Injection 5.1 pH Evolution The simulation results from the CO2 only case are similar to those from CO2 + H2S. The results from CO2 + SO2 case are however, significantly different (Figure 5). CO2 and CO2 + H2S injection leads to lower pH (~4.5, less so comparing to siliciclastic reservoir) near the well bore, resulting from the dissolution of CO2 and H2S in the formation water and quicker buffering reactions. Once the injection stops, the pH increases slightly (~5) but does not rebound to more neutral pH due to limited buffering capability. Co-injection of SO2 with CO2 leads to very low pH (~0), presumably due to the dissolution of SO2 and the formation of a sulphuric acid zone close to the injection well. Corrosion and well abandonment can be potential issues. After the injection, the pH is buffered but still remains considerably low near the well bore. 8.0
8.0
7.5
7.0
1,000 yr
7.0
6.0
10,000 yr
6.5
5.0
100 yr 7.0
6.0
pH
pH
6.5
pH
8.0
10 yr
7.5
6.0
4.0
5.5
5.5
3.0
5.0
5.0
2.0
4.5
4.5
1.0
4.0
4.0
0
1
10
100
Radial Distance (m)
CO2 Only
1000
10000
pH ~ 0!
0.0 0
1
10
100
Radial Distance (m)
CO2 + H2S
1000
10000
0
1
10
100
1000
10000
Radial Distance (m)
CO2 + SO2
Figure 5. Simulated pH distribution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
1788
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
6
Xiao et al / Energy Procedia 00 (2008) 000–000
5.2 Mineral Dissolution/Precipitation Simulation results reveal that calcite dissolution occurs in all cases at a much faster rate compared to the siliciclastic reservoir injection. Under mixed CO2 + SO2 injection, significant and rapid calcite dissolution and anhydrite precipitation occurs near the well bore with a sharp reaction front. The latter could lead to significant loss of injection performance due to porosity reduction. The simulation results are comparable with field observations of acid gas injections in carbonate reservoirs [10]. 0.8
0.8
0.7
0.7
0 yr
0.4
10 yr
0.3
100 yr
0.2
1,000 yr 10,000 yr
0.1 0.0
0.6
Calcite (vol %)
Calcite (vol %)
0.5
0.5 0.4
Calcite Dissolution
0.3
1
10
100
1000
0.4 0.3 0.2
0.1
0.1
10000
0
Radial Distance (m)
1
10
100
1000
10000
0.0 0
Radial Distance (m)
CO2 Only
0 yr
0.5
0.2
0.0 0
Calcite Dissolution Anhydrite Precipitation
0.7
0 yr
0.6
0.6
Calcite (vol %)
0.8
1
10
100
1000
10000
Radial Distance (m)
CO2 + H2S
CO2 + SO2
Figure 6. Simulated mineral (calcite) evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
6. Porosity Evolution Siliciclastic Reservoir: The gas injection leads to an increase in porosity close to the well due to net mineral dissolution, and a decrease at distances due to mineral trapping in all three cases (Figure 7). However, the porosity gains associated with CO2 and CO2 + H2S injection are much smaller (0.30 to 0.32) compared to the CO2 + SO2 injection (0.30 to 0.50) at 100 years. The same trend is observed for mineral trapping. 0.6
10 yr
0.6
0.6
0.5
0.5
Dissolution
1,000 yr 10,000 yr
0.4
Dissolution
0.3
Porosity
Porosity
0.5
0.4
Dissolution
Trapping 0.2
0
10 Radial Distance (m)
CO2 Only
1000
Trapping 0 yr
0 yr
Trapping 0.2
0.4
0.3
0.3
0 yr
Porosity
100 yr
0
10 Radial Distance (m)
CO2 + H2S
1000
0.2 0
10
1000
Radial Distance (m)
CO2 + SO2
Figure 7. Simulated porosity evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a siliciclastic reservoir (after Xu et al [1]).
Carbonate Reservoir: Simulation results indicate that there are significant porosity increases (0.30 to 0.40 at 100 years) associated with CO2 and CO2 + H2S injection due to calcite dissolution near the well bore (Figure 8).
1789
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790 Xiao et al / Energy Procedia 00 (2008) 000–000
7
However, there is significant porosity decrease (0.30 to 0.20) associated with CO2 + SO2 injection due to anhydrite precipitation. There is little mineral trapping in all three mixed gas injection scenarios, suggesting limited CO2 sequestration capacity in a limestone dominated carbonate reservoir. 0.45 0.40
Calcite Dissolution
0.40
0.35
0.25
0 yr
100 yr 1,000 yr
0.20
0.30
0 yr
0
10 1000 Radial Distance (m)
CO2 Only
0.30
0 yr
0.25
0.25
0.20
0.20
0.15
0.15
10,000 yr 0.15
Porosity
10 yr
Porosity
0.30
Calcite Dissolution Anhydrite Precipitation
0.40
0.35
0.35 Porosity
0.45
0.45
Calcite Dissolution
0
10 1000 Radial Distance (m)
CO2 + H2S
Porosity Decrease 0
10 1000 Radial Distance (m)
CO2 + SO2
Figure 8. Simulated porosity evolution as a function of radial distance at 10, 100, 1,000, and 10,000 years for the three mixed gas injection scenarios in a carbonate reservoir.
7. Summary and Conclusions x x x x
x
The behavior of CO2 and acid gas injection and storage is controlled by gas mixtures, reservoir mineralogy, timing, and injection design. Co-injection of H2S yields similar behavior compared to CO2 injection. Co-injection of SO2 results in the formation of sulfuric acid close to the well. Corrosion of pipes and well abandonment are potential issues. Significantly more CO2 is sequestered as ankerite and dawsonite in siliciclastic reservoirs than in carbonate reservoirs, with the mineral trapping capability reaching as much as 80 kg/m3. Most injection scenarios result in porosity increase close to the well and decrease at distances. However, co-injection of SO2 in a carbonate reservoir leads to significant anhydrite precipitation and porosity reduction in the near-well region. The results from Reactive Transport Modeling provide valuable insights for describing, analyzing, interpreting, and assessing the physical properties and dynamic behaviours of injected CO2 and facilitating the screening and evaluation of CO2 storage strategy.
8. Acknowledgments The paper benefited from fruitful discussions with ExxonMobil colleagues Will Maze, Gary Teletzke, John Wilkinson, Lee Esch, Bob Pottorf, and Mark Richardson. Careful reviews by Chris Reaves and Bill Clendenen improved the manuscript. Permission by ExxonMobil to release the paper is greatly appreciated. Drs. Tianfu Xu and Karsten Pruess were supported by the Zero Emission Research and Technology project (ZERT) of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 with Lawrence Berkeley National Laboratory.
1790
Y. Xiao et al. / Energy Procedia 1 (2009) 1783–1790
8
Xiao et al / Energy Procedia 00 (2008) 000–000
9. References 1.
Xu, T, J. A. Apps, K. Pruess, and H. Yamamoto, 2007. Numerical modeling of injection and mineral trapping of CO2 with H2S and SO2 in a sandstone formation. Chemical Geology, Volume 242, Issues 3-4, 15 August 2007, Pages 319-346
2.
Knauss, K.G., Johnson, J.W., Steefel, C.I., 2005. Evaluation of the impact of CO2, co-contaminant gas, aqueous fluid, and reservoir rock interactions on the geologic sequestration of CO2. Chem. Geol. 217, 339350.
3.
Pruess, K., Xu, T., Apps, J., García, J., 2003. Numerical modeling of aquifer disposal of CO2. SPE (Society of Petroleum Engineers) Journal, p. 49-59, March (2003).
4.
Johnson, J.W., Nitao, J.J., Steefel, C.I., Knaus, K.G., 2001. Reactive transport modeling of geologic CO2 sequestration in saline aquifers: The influence of intra-aquifer shales and the relative effectiveness of structural, solubility, and mineral trapping during prograde and retrograde sequestration, In proceedings: First National Conference on Carbon Sequestration. Washington, DC.
5.
Arvidson, R. S. and Mackenzie, F.T. 1999. The dolomite problem: control of precipitation kinetics by temperature and saturation state: American Journal of Science, v. 299, p. 257-288.
6.
Xu, T., E.L. Sonnenthal, N. Spycher, and K. Pruess, 2006. TOURGHREACT: A simulation program for non-isothermal multiphase reactive geochemical transport in variably saturated geologic media. Computer & Geosciences 32, 145-165.
7.
Watson, M.N., Zwingmann, N., Lemon, N.M., 2004. The Ladbroke Grove–Katnook carbon dioxide natural laboratory: A recent CO2 accumulation in a lithic sandstone reservoir. Energy, 29(9-10), 1457-1466.
8.
Moore, J., Adams, M., Allis, R., Lutz, S., Rauzi, S., 2005. Mineralogical and geochemical consequences of the long-term presence of CO2 in natural reservoirs: An example from the Springerville–St. Johns Field, Arizona, and New Mexico, U.S.A. Chemical Geology, 217, 365-385.
9.
Worden R H (2006) Dawsonite cement in the Triassic Lam Formation, Shabwa Basin, Yemen: a natural analogue for a potential mineral product of subsurface CO2 storage for greenhouse gas reduction. Marine and Petroleum Geology, 23, pp.61-67.
10. Huang, N S, G.E. Aho, B.H. Baker and P.A. Foglesong, T.R. Matthews, R.J. Pottorf, 2007. Integrated Reservoir Modeling to Maximize the Value of a Large Sour-Gas Field with High Concentrations of Inerts. 2007 International Petroleum Technology Conference, Dubai, UAE.